Power converters generally include switches and one or more capacitors, for example, to power portable electronic devices and consumer electronics. A switch-mode power converter is a specific type of power converter that regulates its output voltage or current by switching storage elements (i.e. inductors and capacitors) into different electrical configurations using a switch network.
One type of switch-mode power converter is the switched capacitor converter. A switched capacitor converter uses capacitors to transfer energy. As the transformation ratio increases, the number of capacitors and switches increases.
A switch capacitor converter includes a switch network containing numerous switches. These switches are active devices that are usually implemented with transistors. The switch network can be integrated on a single or on multiple monolithic semiconductor substrates.
Typical power converters perform voltage transformation and output regulation. In many power converters, such as a buck converter, this is carried out in a single stage. However, it is also possible to split these two functions into two specialized stages. Such two-stage power converter architectures feature a transformation stage and a separate regulation stage. The transformation stage transforms one voltage into another, while the regulation stage ensures that the voltage and/or current output of the transformation stage maintains desired characteristics.
An example of a two-stage power converter architecture is illustrated in FIG. 1A, where capacitors are utilized to transfer energy. The transformation stage is represented by a switched-capacitor element 12A, which closely resembles a switched capacitor converter while the regulation stage is represented by a regulating circuit 16A.
In this architecture, a switched capacitor element 12A is electrically connected to a voltage source 14 at an input end thereof. An input of a regulating circuit 16A is electrically connected to an output of the switched capacitor element 12A. A load 18A is then electrically connected to an output of the regulating circuit 16A. Such a converter is described in US Patent Publication 2009/0278520, filed on May 8, 2009, the contents of which are herein incorporated by reference. Furthermore, a modular multi-stage power converter architecture was described in PCT Application PCT/2012/36455, filed on May 4, 2012, the contents of which are also incorporated herein by reference.
The switched capacitor element 12A and regulating circuit 16A can be mixed and matched in a variety of different ways. This provides a transformative integrated power solution (TIPS™) for the assembly of such converters. As such, the configuration shown in FIG. 1A represents only one of multiple ways to configure one or more switched capacitor elements 12 with one or more regulating circuits 16A.
Typically, the switch network of the switched capacitor element 12A and the regulating circuit 16A are fabricated in a semiconductor process that has passive devices. However, these passive devices are normally used in the analog circuitry to control the power converter. They are not normally used to store energy in the power converter. This is because these passive devices cannot efficiently store a significant amount of energy.
These passive devices are usually planar and fabricated after the active devices in a higher level of metal to reduce parasitic effects. Since these passive devices are fabricated after the active devices, and on the same wafer as the active devices, the processing steps for making these passive devices should be chosen carefully. An incorrect choice may damage the active devices that have already been fabricated.
To avoid possibly damaging the active devices during fabrication of the passive devices, it is preferable to only use CMOS compatible processing. Given this processing requirement, it is difficult and/or expensive to achieve high capacitance density capacitors or high Q inductors in a CMOS flow. Therefore, in power converters, it is common practice to store energy in discrete components, such as multilayer ceramic capacitors and chip inductors. However, it is possible to produce inexpensive high performance passive devices in their own wafer and process flow that can be used in specific applications. These devices will be referred to as integrated passive devices (IPDs).
An implementation of the power converter architecture shown in FIG. 1A is illustrated in FIG. 1B-1D.
In the embodiment shown in FIG. 1B, a power converter 20 draws energy from a voltage source 14 at a high input voltage VIN and delivers that energy to a load 18A at a low output voltage VO. Without loss of generality, the load 18A is modeled as a resistor.
The power converter 20 includes a switched capacitor element 12A that features a 3:1 series-parallel switched capacitor network having power switches S1-S7 and pump capacitors C21-C22. In contrast, the regulating circuit 16A is a buck converter having first and second output power switches SL, SH, a filter inductor L1, and an output capacitor CO. The power switches S1-S7, the output power switches SL, SH, and the driver/control circuitry 23 are integrated in a single semiconductor die 22. However, the pump capacitors C21-C22, the filter inductor L1, and a decoupling input capacitor CIN1 are discrete components.
In operation, the power switches S1, S3, S6 and the power switches S2, S4, S5, S7 are always in complementary states. Thus, in a first switch state, the power switches S1, S3, S6 are open and the power switches S2, S4, S5, S7 are closed. In a second switch state, the power switches S1, S3, S6 are closed and the power switches S2, S4, S5, S7 are open. Similarly, the output power switches SL, SH are in complementary states.
Typically, the regulating circuit 16A operates at higher switching frequencies than the switched capacitor element 12A. However, there is no requirement of any particular relationship between the switching frequencies of the regulating circuit 16A and the switching frequency of the switched capacitor element 12A. The driver/control circuitry 23 provides the necessary power to activate the switches and controls the proper switch states to ensure a regulated output voltage VO.
In power converters, it is common practice to solder a semiconductor die 22 or packaged die to an electrical interface 28, and to then horizontally mount capacitors and inductors on the electrical interface 28 around the semiconductor die 22. Such an arrangement is shown in a top view in FIG. 1D and in a side view in FIG. 1C taken along a line 24 in FIG. 1D.
An electrical interface 28 provides electrical conductivity between the power converter 20 and a load to which the power converter 20 is ultimately supplying power. Examples of electrical interfaces 28 include printed circuit boards, package lead frames, and high density laminates.
The discrete components in the power converter 20 include the pump capacitors C21-C22, the input capacitor CIN1, the output capacitor CO, and the filter inductor L1. These discrete components are horizontally disposed with respect to the semiconductor die 22 and electrically coupled to the die 22 bytraces on the electrical interface 28.
Each power switch in the power converter 20 is typically composed of numerous smaller switches connected in parallel as illustrated by the close-up 26 in FIG. 1D. This allows the power switches to carry a large amount of current without overheating.